A method, system and storage medium for calculating available power transmission capacity of a power system

Abstract

The present application relates to the technical field of available transmission capacity calculation, in particular to a power system available transmission capacity calculation method, system and storage medium; the operation data and the grid topology structure data of the power system are acquired, and the initial power flow model of the power system is constructed; the photovoltaic power generation prediction model is constructed to identify the meteorological data and the historical power generation data, the photovoltaic injection power of the photovoltaic power station to the power system in the future period is predicted; based on the operation data and the grid topology structure data, the transmission path of the photovoltaic injection power in the power system is identified, and the dynamic path loss is calculated based on the transmission path; the initial power flow model is updated based on the photovoltaic injection power and the dynamic path loss, and the prediction power flow model is obtained; under the preset safety and stability constraint condition, the maximum power transmission limit of the prediction power flow model is solved. The present application accurately identifies the available transmission capacity through the maximum power transmission limit.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of available transmission capacity calculation, in particular to a method and system for calculating available transmission capacity of a power system and a storage medium. BACKGROUND

[0002] With the transformation of global energy structure, the proportion of renewable energy represented by photovoltaic in the power system is increasing. However, photovoltaic power generation has significant intermittency, volatility and uncertainty, which brings great challenges to the safe and stable operation of the power system. Available transmission capacity is a measure of the effective power that can be transmitted on the transmission section under the premise of meeting the safety and stability constraints, and is a key basis for power market transactions and grid dispatching decisions.

[0003] Traditional available transmission capacity calculation methods are mostly based on static and deterministic grid models and load forecasts, which are difficult to adapt to modern power systems with high proportion of photovoltaic. The main defects are as follows: the traditional method usually uses a simplified photovoltaic prediction model, which cannot accurately depict the dramatic power fluctuations caused by cloud cover and other factors, and also ignores the dynamic changes of the health status of photovoltaic devices, resulting in large prediction errors. At the same time, the parameters such as line impedance in the grid model usually use offline set typical values, which cannot reflect the actual dynamic changes caused by factors such as temperature and load rate, resulting in the calculated power flow results not consistent with the actual situation, and further affecting the accuracy of available transmission capacity calculation.

[0004] Therefore, how to establish a dynamic model that can accurately reflect the future system state under the background of high proportion of photovoltaic access, and on this basis, realize the rapid and accurate calculation of available transmission capacity, is a technical problem in the field of power systems.

[0005] Therefore, a method and system for calculating available transmission capacity of a power system and a storage medium are proposed. SUMMARY

[0006] The purpose of the present application is to provide a method and system for calculating available transmission capacity of a power system and a storage medium, which accurately identifies the available transmission capacity by solving the maximum power transmission limit of the predicted power flow model under the preset safety and stability constraints.

[0007] To achieve the above purpose, the present application provides the following technical solutions:

[0008] A method for calculating available transmission capacity of a power system, comprising:

[0009] Obtaining operation data and grid topology structure data of the power system to construct an initial power flow model of the power system;

[0010] The photovoltaic power generation prediction model is based on meteorological data and historical power generation data, and predicts the photovoltaic injection power of the photovoltaic power station to the power system in a future period;

[0011] Based on the operation data and the power grid topology data, the transmission path of the photovoltaic injection power in the power system is identified, and the dynamic path loss is calculated based on the transmission path;

[0012] Based on the photovoltaic injection power and the dynamic path loss, the initial power flow model is updated to obtain a predicted power flow model;

[0013] Under the preset safety and stability constraint condition, the maximum power transmission limit of the predicted power flow model is solved to determine the available transmission capacity of the power system.

[0014] The photovoltaic power generation prediction model comprises a dynamic health identification layer, a space-time disturbance identification layer, and a photovoltaic power generation prediction layer;

[0015] The dynamic health identification layer identifies the aging influence parameter based on the factory service life of the photovoltaic power generation device and the historical power generation data, and identifies the dust accumulation influence parameter of the photovoltaic power generation device based on the rainfall data, the atmospheric suspended particulate matter concentration and the meteorological data;

[0016] The space-time disturbance identification layer extracts the shape feature, thickness feature, speed feature and direction feature of the cloud layer through a convolutional neural network to identify the space-time interference parameter of the cloud layer to the photovoltaic power generation, wherein the space-time interference parameter comprises interference time data and a space-time interference factor;

[0017] The photovoltaic power generation prediction layer predicts the photovoltaic injection power based on the normal direct radiation, horizontal scattering radiation, ambient temperature and wind speed in the meteorological data, in combination with the aging influence parameter, the dust accumulation influence parameter and the space-time interference parameter of the photovoltaic power generation device.

[0018] The power grid topology data of the power system is obtained through a power network management system;

[0019] The power flow distribution path of the electric energy from the photovoltaic power station to each part of the power grid is analyzed based on the power grid topology data by using a power flow tracking algorithm, as the transmission path;

[0020] The voltage phase angle data and the voltage amplitude data collected by the phasor measurement unit deployed at the power grid node are obtained;

[0021] The line impedance parameter of the transmission path in the power flow tracking algorithm is corrected using the voltage phase angle data and the voltage amplitude data, and the dynamic path loss is calculated based on the corrected line impedance parameter.

[0022] The process of correcting the line impedance parameter comprises:

[0023] determining transmission paths based on the power flow tracing algorithm, locking the line set which has significant influence on photovoltaic power transmission;

[0024] extracting voltage phase angle data and voltage amplitude data of the line set at both ends and adjacent key nodes;

[0025] constructing an objective function of minimum power flow equation residual based on line impedance as a variable;

[0026] solving the minimized objective function by using the Gauss-Newton method to obtain the correction value of the line impedance.

[0027] Through offline simulation calculation, a multi-dimensional operating state space containing photovoltaic injection power and key section active power is constructed, and a multi-dimensional safety domain hyperbody satisfying the N-1 safety criterion, thermal stability constraint and voltage stability constraint is determined;

[0028] projecting the photovoltaic injection power and the key section active power determined by the predicted power flow model into the multi-dimensional operating state space to obtain a current operating point;

[0029] Starting from the current operating point, the intersection point of the multi-dimensional safety domain hyperbody boundary is calculated along the direction of increasing coordinate axis of the key section active power, and the active power coordinate value corresponding to the intersection point is determined as the total transmission capacity;

[0030] The total transmission capacity is reduced by the active power value of the transmission section corresponding to the current operating point, and further reduced by the preset transmission reliability margin and capacity benefit margin to obtain the available transmission capacity of the power system.

[0031] The key section is the transmission channel with the strongest restriction on photovoltaic injection power transmission capacity obtained by multiple power flow calculations and static security analysis of the power system.

[0032] A system for calculating the available transmission capacity of a power system, comprising:

[0033] The system construction module obtains the operating data and grid topology data of the power system, and constructs an initial power flow model of the power system;

[0034] The power calculation module constructs a photovoltaic power generation prediction model, which is based on meteorological data and historical power generation data to predict the photovoltaic injection power of the photovoltaic power station to the power system in the future period;

[0035] The loss identification module identifies the transmission path of the photovoltaic injection power in the power system based on the operating data and grid topology data, and calculates the dynamic path loss based on the transmission path;

[0036] A model updating module updates the initial power flow model based on the photovoltaic injection power and the dynamic path loss to obtain a predicted power flow model;

[0037] A power transmission identification module solves the maximum power transmission limit of the predicted power flow model under preset safe and stable constraint conditions to determine the available power transmission capacity of the power system.

[0038] A storage medium, the storage medium stores a computer program, the computer program is executed by the processor to realize the above-mentioned one kind of power system available power transmission capacity calculation method.

[0039] Compared with the prior art, the beneficial effects of the present application are:

[0040] 1、The present application identifies the aging influence parameter based on the service life and historical power generation data of the photovoltaic power generation device; identifies the dust accumulation influence parameter of the photovoltaic power generation device based on rainfall data, atmospheric suspended particulate matter concentration and weather data; identifies the space-time interference parameter through cloud image data; and then predicts based on weather data, aging influence parameter, dust accumulation influence parameter and space-time interference parameter to accurately obtain the photovoltaic injection power.

[0041] 2、The present application uses a power flow tracking algorithm to identify grid topology structure data, analyzes the power flow distribution path of the power flow from the photovoltaic power station to each place in the power grid to obtain a transmission path; and then corrects the transmission path line impedance parameter based on the objective function of minimum power flow equation residual error and voltage phase angle data and voltage amplitude data to accurately calculate the dynamic path loss.

[0042] 3、The present application constructs a multi-dimensional operating state space containing photovoltaic injection power and key section active power through offline simulation calculation, and determines a multi-dimensional safety domain super body which satisfies the N-1 safety criterion, thermal stability constraint and voltage stability constraint; the photovoltaic injection power and key section active power determined by the predicted power flow model are projected into the multi-dimensional operating state space to accurately identify the total power transmission capacity. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 It is a flowchart of the calculation method of the available power transmission capacity of the power system of the present application;

[0044] Figure 2 It is a structure diagram of the photovoltaic power generation prediction model of the present application;

[0045] Figure 3 It is a structure diagram of the calculation system of the available power transmission capacity of the power system of the present application. DETAILED DESCRIPTION

[0046] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those ordinarily skilled in the art without creative effort belong to the scope of the present application.

[0047] Embodiment one:

[0048] The present application provides a method for calculating the available transmission capacity of a power system, the flowchart of which is shown in Figure 1 , including:

[0049] Obtain the operation data and the grid topology data of the power system, and construct an initial power flow model of the power system;

[0050] Construct a photovoltaic power generation prediction model, which is based on meteorological data and historical power generation data to predict the photovoltaic injection power of the photovoltaic power station to the power system in a future period;

[0051] Based on the operation data and the grid topology data, identify the transmission path of the photovoltaic injection power in the power system, and calculate the dynamic path loss based on the transmission path;

[0052] Update the initial power flow model based on the photovoltaic injection power and the dynamic path loss to obtain a predicted power flow model;

[0053] Under the preset safety and stability constraint condition, solve the maximum power transmission limit of the predicted power flow model to determine the available transmission capacity of the power system.

[0054] Through the energy management system (EMS) of the regional power grid, obtain the topology information of the current power grid, including the connection relationship and switch state of the bus, line and transformer. Through the wide area measurement system (WAMS) and SCADA system of the power grid, obtain the operation data of each node; the operation data includes the output of the conventional generator, the load consumption power, the voltage phase angle data and the voltage amplitude data, etc. Using the standard power system analysis software, based on the above topology and operation data, a benchmark power flow model reflecting the current power grid operation state is established. The model contains the parameters of all network elements.

[0055] The photovoltaic power generation prediction model includes a dynamic health identification layer, a time and space disturbance identification layer and a photovoltaic power generation prediction layer, and its structure is shown in Figure 2 .

[0056] The dynamic health identification layer identifies an aging influence parameter based on a factory use time and historical power generation data of the photovoltaic power generation device; and identifies a dust accumulation influence parameter of the photovoltaic power generation device based on rainfall data, atmospheric suspended particulate matter concentration and meteorological data.

[0057] The space-time disturbance identification layer extracts shape features, thickness features, speed features and direction features of the cloud layer through a convolutional neural network, and identifies a space-time disturbance parameter of the cloud layer on the photovoltaic power generation, the space-time disturbance parameter including disturbance time data and a space-time disturbance factor.

[0058] The photovoltaic power generation prediction layer predicts the photovoltaic injection power based on normal direct radiation, horizontal scattering radiation, ambient temperature and wind speed in the meteorological data, in combination with the aging influence parameter, the dust accumulation influence parameter and the space-time disturbance parameter of the photovoltaic power generation device.

[0059] Preferably, the aging influence parameter reflects material aging influence of the photovoltaic power generation device, and is identified according to the factory use time and historical power generation data, the historical power generation data including historical meteorological data and power generation efficiency. The power generation efficiency specifically refers to an efficiency of the photovoltaic power generation device in converting received solar energy into electric energy, and the photovoltaic power generation device includes a photovoltaic power generation panel.

[0060] Preferably, the dust accumulation influence parameter is used to reflect influence of surface dust of the photovoltaic power generation device on its power generation efficiency, and is predicted according to atmospheric suspended particulate matter concentration data and meteorological data in a cumulative no-rainfall time period.

[0061] Further, the system continuously tracks data such as "cumulative no-rainfall time" and "atmospheric suspended particulate matter concentration" of each photovoltaic power station. In combination with the attenuation of the power generation efficiency, a dynamic dust accumulation index is calculated and displayed, the dust accumulation index being obtained based on power generation loss caused by dust accumulation and cost of cleaning and maintenance. When the index exceeds a preset threshold value, for example, daily power generation loss caused by dust accumulation has exceeded the cost of cleaning once, the system will automatically generate an operation and maintenance suggestion; after manual cleaning, the "cumulative no-rainfall time" is re-determined. Through dynamic measurement of the loss of photovoltaic power generation effect and the cleaning cost, manual cleaning can be planned based on the goal of optimal benefit.

[0062] The space-time disturbance parameter acquisition process includes:

[0063] Cloud image data of a region where the photovoltaic power generation device is located is acquired, a 3D-CNN structure is adopted, a core structure of which includes three Conv3D layers for extracting space-time joint features, a MaxPooling3D layer is connected after the core structure for dimension reduction, and finally a Flatten layer and two fully connected layers are used.

[0064] Based on shape feature, speed feature and direction feature recognition, combined with the location of the photovoltaic power generation device or photovoltaic power plant, interference time data is recognized; the interference time data is the time period data of cloud layer interference; the spatio-temporal interference factor is recognized according to the thickness feature, and the spatio-temporal interference factor reflects the influence of cloud layer thickness on the power generation efficiency of the photovoltaic power generation device.

[0065] The training process of the spatio-temporal disturbance correction layer includes: using historical cloud image data and corresponding photovoltaic power station measured efficiency data for supervised training. The input is a cloud image sequence, and the label is the measured efficiency / theoretical clear sky efficiency. The Adam optimizer and the mean square error (MSE) loss function are used for iterative optimization of the model parameters.

[0066] Preferably, the meteorological data includes cloud image data, normal direct radiation, horizontal scattering radiation, ambient temperature, and wind speed; the normal direct radiation refers to the solar radiation intensity perpendicular to the direction of sunlight; and the horizontal scattering radiation refers to the solar radiation intensity projected on the horizontal plane from the sky scattering.

[0067] The reference power generation prediction layer is constructed based on a neural network model and is trained by labeled meteorological data, aging influence parameters, dust accumulation influence parameters and spatio-temporal disturbance parameters.

[0068] Further, the photovoltaic power station and the energy storage are dynamically optimized based on the photovoltaic injection power:

[0069] The spatio-temporal disturbance layer is used to predict the cloud change in the future period, so as to plan the optimal charging and discharging path for the energy storage system matched with the photovoltaic power station. If the model predicts that there will be a large piece of cloud layer passing through, resulting in reduced power generation, the system will instruct the energy storage system to reduce charging or even discharge in advance to smooth the power generation gap. On the contrary, if a long period of sunny weather is predicted, the energy storage system will be charged during this period.

[0070] By discharging in the power generation valley and charging in the power generation peak, the economic benefit is maximized; the operation mode of the energy storage is upgraded from passive response based on the current power to active management based on future prediction; not only the power can be better smoothed, but also the efficiency of the energy storage system can be improved through low charging and high discharging.

[0071] The aging influence parameters are recognized based on the factory use life and historical power generation data of the photovoltaic power generation device; the dust accumulation influence parameters of the photovoltaic power generation device are recognized based on rainfall data, atmospheric suspended particulate matter concentration and meteorological data; the spatio-temporal disturbance parameters are recognized through cloud image data; and the photovoltaic injection power is accurately obtained based on the meteorological data, the aging influence parameters, the dust accumulation influence parameters and the spatio-temporal disturbance parameters.

[0072] The power grid topology structure data of the power system is obtained through the power network management system;

[0073] A power flow tracking algorithm is used to analyze the power flow distribution path of the power generated by the photovoltaic power station to each part of the power grid based on the power grid topology data, as the transmission path;

[0074] Voltage phase angle data and voltage amplitude data collected by a phasor measurement unit deployed at a power grid node are obtained;

[0075] The line impedance parameters of the transmission path in the power flow tracking algorithm are corrected using the voltage phase angle data and the voltage amplitude data, and the dynamic path loss is calculated based on the corrected line impedance parameters.

[0076] The process of correcting the line impedance parameters includes:

[0077] Based on the transmission path determined by the power flow tracking algorithm, a line set that significantly affects photovoltaic power transmission is locked;

[0078] Voltage phase angle data and voltage amplitude data of the first and last ends of the line set and adjacent key nodes are extracted;

[0079] A target function of minimizing the residual error of the power flow equation with the line impedance as the variable to be solved is constructed;

[0080] The Gauss-Newton method is used to solve the minimized target function to obtain the corrected value of the line impedance.

[0081] The present application uses a power flow tracking algorithm to identify power grid topology data, analyzes the power flow distribution path of the power generated by the photovoltaic power station to each part of the power grid, and obtains the transmission path; based on the target function of minimizing the residual error of the power flow equation and the voltage phase angle data and the voltage amplitude data, the line impedance parameters of the transmission path are corrected, and the dynamic path loss is accurately calculated.

[0082] Through offline simulation calculation, a multi-dimensional operating state space containing photovoltaic injection power and key section active power is constructed, and a multi-dimensional safety domain hyperbody that meets the N-1 safety criterion, thermal stability constraint and voltage stability constraint is determined;

[0083] The photovoltaic injection power and the key section active power determined by the predicted power flow model are projected into the multi-dimensional operating state space to obtain the current operating point;

[0084] Starting from the current operating point, the intersection point of the multi-dimensional safety domain hyperbody boundary is calculated along the direction of growth of the key section active power dimension coordinate axis, and the active power coordinate value corresponding to the intersection point is determined as the total power transmission capacity;

[0085] The total transmission capability is subtracted by the active power value of the transmission section corresponding to the current operating point, and further subtracted by the preset transmission reliability margin and capacity benefit margin, to obtain the available transmission capacity of the power system.

[0086] The section in the power system refers to a set of transmission lines, which together form a power transmission channel connecting two regions. The critical section is the transmission channel with the strongest limitation on photovoltaic injection power transmission capacity obtained by multiple power flow calculations and static security analysis of the power system.

[0087] The N-1 safety criterion is an industry standard for power system planning and operation, which requires that after the sudden tripping of any single important device (any one of N devices, such as a line or a transformer) in the grid due to failure, the remaining N-1 devices can still operate stably and cannot cause cascading reactions such as device overload and large-scale power outage.

[0088] Thermal stability constraint refers to the heat generated by current flowing through the conductor. If the transmitted power is too large, the conductor will expand and stretch due to overheating, resulting in insufficient distance to the ground, and even burning out. This constraint requires that at any time, the temperature (or load current) of all lines cannot exceed the design upper limit.

[0089] Voltage stability constraint: For the normal operation of the power grid, the voltage of all nodes must be maintained at a relatively stable level; if the transmitted power is too large or the line is too long, it may cause the end voltage to be too low, and voltage that is too low or collapses will cause user appliances to malfunction or even be damaged.

[0090] Transmission reliability margin is the capacity reserved to cope with uncertainties, and is the capacity that must be reserved to ensure reliable operation of the system in the event of unexpected situations. It is usually set according to experience or probability statistical methods.

[0091] Capacity benefit margin is the capacity reserved to ensure fair and efficient power markets, ensuring that the transmission network can provide services for temporary power transactions that do not have long-term contracts, so that more power generators can participate in market competition and improve the economic efficiency of the entire power system.

[0092] Available transmission capacity is the real available transmission capacity that we finally announce to the market and dispatchers, which can be used to arrange new power transactions or accept new power sources. Its calculation formula is: available transmission capacity = total transmission capability - transmission reliability margin - capacity benefit margin.

[0093] The application constructs a multi-dimensional operation state space containing photovoltaic injection power and key section active power through offline simulation calculation, and determines a multi-dimensional safety domain super body meeting N-1 safety criterion, thermal stability constraint and voltage stability constraint; the photovoltaic injection power and the key section active power determined by the power flow prediction model are projected into the multi-dimensional operation state space, and the total power transmission capacity is accurately identified.

[0094] Further, in the original photovoltaic prediction model, the Quantile Regression or Monte Carlo Dropout technology is used in the neural network, so that the output is no longer a single power prediction value, but a probability interval containing future power distribution; the uncertainty of photovoltaic output is propagated to the power flow of the key section through power flow calculation, so that the finally calculated total power transmission capacity and available power transmission capacity also become a probability distribution. The grid dispatching decision can be changed from traditional deterministic margin management to more refined risk management, and the power transmission channel and renewable energy are maximized under the premise of ensuring safety.

[0095] The application also provides a power system available power transmission capacity calculation system, which has the structure as shown in the accompanying drawings, and comprises: Figure 3

[0096] A system construction module acquires operation data and grid topology structure data of the power system, and constructs an initial power flow model of the power system;

[0097] A power calculation module constructs a photovoltaic power generation prediction model, the photovoltaic power generation prediction model predicts photovoltaic injection power of the photovoltaic power station to the power system in a future period based on meteorological data and historical power generation data;

[0098] A loss identification module identifies a transmission path of the photovoltaic injection power in the power system based on the operation data and the grid topology structure data, and calculates a dynamic path loss based on the transmission path;

[0099] A model updating module updates the initial power flow model based on the photovoltaic injection power and the dynamic path loss, and obtains a power flow prediction model;

[0100] A power transmission identification module determines the available power transmission capacity of the power system by solving a maximum power transmission limit of the power flow prediction model under preset safety and stability constraints.

[0101] ​The application is based on operation data and grid topology data of a power system to construct an initial power flow model of the power system; a photovoltaic power generation prediction model is constructed to identify meteorological data and historical power generation data, predict photovoltaic injection power of a photovoltaic power station to the power system in a future period, identify transmission paths of the photovoltaic injection power in the power system based on the operation data and the grid topology data, and calculate dynamic path losses based on the transmission paths; the initial power flow model is updated based on the photovoltaic injection power and the dynamic path losses to obtain a predicted power flow model; and the maximum power transmission limit of the predicted power flow model is solved under preset safety and stability constraints to accurately identify available transmission capacity.

[0102] The application provides a storage medium, and a computer program is stored on the storage medium.

[0103] Embodiment two:

[0104] The application provides a method for calculating available transmission capacity of a power system, which comprises the following steps:

[0105] Operation data and grid topology data of the power system are acquired to construct an initial power flow model of the power system;

[0106] A photovoltaic power generation prediction model is constructed, and the photovoltaic power generation prediction model is based on meteorological data and historical power generation data to predict photovoltaic injection power of a photovoltaic power station to the power system in a future period;

[0107] Transmission paths of the photovoltaic injection power in the power system are identified based on the operation data and the grid topology data, and dynamic path losses are calculated based on the transmission paths;

[0108] The initial power flow model is updated based on the photovoltaic injection power and the dynamic path losses to obtain a predicted power flow model;

[0109] The maximum power transmission limit of the predicted power flow model is solved under preset safety and stability constraints to determine the available transmission capacity of the power system.

[0110] The photovoltaic power generation prediction model comprises a dynamic health identification layer, a time-space disturbance identification layer and a photovoltaic power generation prediction layer;

[0111] The dynamic health identification layer identifies an aging influence parameter based on a factory use life of a photovoltaic power generation device and historical power generation data, and identifies a dust accumulation influence parameter of the photovoltaic power generation device based on rainfall data, atmospheric suspended particulate matter concentration and meteorological data;

[0112] The spatio-temporal disturbance identification layer extracts shape features, thickness features, speed features and direction features of the cloud layer through a convolutional neural network to identify spatio-temporal disturbance parameters of the cloud layer on photovoltaic power generation, the spatio-temporal disturbance parameters including disturbance time data and a spatio-temporal disturbance factor.

[0113] The photovoltaic power generation prediction layer predicts based on normal direct radiation, horizontal scattering radiation, ambient temperature and wind speed in meteorological data, in combination with aging influence parameters, dust accumulation influence parameters and spatio-temporal disturbance parameters of the photovoltaic power generation device to obtain photovoltaic injection power.

[0114] Preferably, the aging influence parameters reflect material aging influence of the photovoltaic power generation device and are identified according to factory use time and historical power generation data, the historical power generation data including historical meteorological data and power generation efficiency.

[0115] Preferably, the dust accumulation influence parameters are used to reflect influence of surface dust accumulation of the photovoltaic power generation device on its power generation efficiency, and the dust accumulation influence parameters are predicted based on atmospheric suspended particulate matter concentration data and meteorological data in a cumulative no-rain period.

[0116] Further, the system continuously tracks data such as “cumulative no-rain time” and “atmospheric suspended particulate matter concentration” of each photovoltaic power station. In combination with the attenuation of power generation efficiency, a dynamic dust accumulation index is calculated and displayed, the dust accumulation index being obtained based on power generation loss caused by dust accumulation and cost of cleaning and maintenance. When the index exceeds a preset threshold, for example, daily power generation loss caused by dust accumulation has exceeded the cost of cleaning once, the system will automatically generate an operation and maintenance suggestion.

[0117] The acquisition process of the spatio-temporal disturbance parameters includes:

[0118] Cloud image data of an area where the photovoltaic power generation device is located is acquired, a 3D-CNN structure is used, a core structure of which includes three Conv3D layers for extracting spatio-temporal joint features, a MaxPooling3D layer is connected after the core structure for dimension reduction, and finally a Flatten layer and two fully connected layers are used.

[0119] Based on shape feature recognition, speed feature recognition and direction feature recognition, in combination with the location of the photovoltaic power generation device or photovoltaic power plant, disturbance time data is identified; the disturbance time data is time period data of cloud layer disturbance; and based on thickness feature recognition, a spatio-temporal disturbance factor is identified, the spatio-temporal disturbance factor reflecting influence of cloud layer thickness on power generation efficiency of the photovoltaic power generation device.

[0120] The training process of the spatio-temporal disturbance correction layer includes: using historical cloud image data and corresponding measured efficiency data of photovoltaic power stations for supervised training. The input is a cloud image sequence, and the label is measured efficiency / theoretical clear sky efficiency. An Adam optimizer and a mean square error (MSE) loss function are used for iterative optimization of model parameters.

[0121] Preferably, the meteorological data includes cloud image data, normal direct radiation, horizontal scattering radiation, ambient temperature, wind speed; the normal direct radiation refers to the solar radiation intensity perpendicular to the direction of sunlight; the horizontal scattering radiation refers to the solar radiation intensity projected on the horizontal plane from the sky scattering.

[0122] The reference power generation prediction layer is constructed based on a neural network model and is trained based on labeled meteorological data, aging influence parameters, dust accumulation influence parameters and space-time interference parameters.

[0123] The application identifies the aging influence parameters based on the service life and historical power generation data of the photovoltaic power generation device, identifies the dust accumulation influence parameters of the photovoltaic power generation device based on rainfall data, atmospheric suspended particulate matter concentration and meteorological data, identifies the space-time interference parameters through cloud image data, and then predicts based on the meteorological data, the aging influence parameters, the dust accumulation influence parameters and the space-time interference parameters to accurately obtain the photovoltaic injection power.

[0124] The power grid topology data of the power system is obtained through the power network management system;

[0125] The power flow distribution path of the electric energy emitted by the photovoltaic power station to each place in the power grid is analyzed based on the power grid topology data by using the power flow tracing algorithm, as the transmission path;

[0126] The voltage phase angle data and voltage amplitude data collected by the phasor measurement unit deployed at the power grid node are obtained;

[0127] The line impedance parameters of the transmission path in the power flow tracing algorithm are corrected by using the voltage phase angle data and the voltage amplitude data, and the dynamic path loss is calculated based on the corrected line impedance parameters.

[0128] The process of correcting the line impedance parameters includes:

[0129] Based on the transmission path determined by the power flow tracing algorithm, the line set which has a significant influence on photovoltaic power transmission is locked;

[0130] The voltage phase angle data and voltage amplitude data of the first and last ends of the line set and the adjacent key nodes are extracted;

[0131] A target function of minimizing the residual error of the power flow equation with the line impedance as the variable is constructed;

[0132] The Gauss-Newton method is used to solve the minimized target function to obtain the correction value of the line impedance.

[0133] The application utilizes a power flow tracking algorithm to identify power grid topological structure data, analyzes a power flow distribution path of the power emitted by the photovoltaic power station to each part of the power grid to obtain a transmission path, and corrects the transmission path line impedance parameters based on a target function of minimum residual error of the power flow equation and voltage phase angle data and voltage amplitude data to accurately measure and calculate a dynamic path loss.

[0134] By offline simulation calculation, a multi-dimensional operating state space containing photovoltaic injection power and key section active power is constructed, and a multi-dimensional safety domain super body satisfying N-1 safety criteria, thermal stability constraints and voltage stability constraints is determined;

[0135] The photovoltaic injection power and the key section active power determined by the predicted power flow model are projected into the multi-dimensional operating state space to obtain a current operating point.

[0136] Starting from the current operating point, the intersection point of the multi-dimensional safety domain super body boundary is calculated along the direction of growth of the key section active power dimension coordinate axis, and the active power coordinate value corresponding to the intersection point is determined as the total power transmission capacity.

[0137] The total power transmission capacity is subtracted by the active power value of the power transmission section corresponding to the current operating point, and further subtracted by a preset power transmission reliability margin and a capacity benefit margin to obtain the available power transmission capacity of the power system.

[0138] The key section is the transmission channel with the strongest photovoltaic injection power transmission capacity limitation obtained by multiple power flow calculations and static security analysis of the power system.

[0139] The application constructs a multi-dimensional operating state space containing photovoltaic injection power and key section active power by offline simulation calculation, and determines a multi-dimensional safety domain super body satisfying N-1 safety criteria, thermal stability constraints and voltage stability constraints; the photovoltaic injection power and the key section active power determined by the predicted power flow model are projected into the multi-dimensional operating state space to accurately identify the total power transmission capacity.

[0140] Although the embodiments of the application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the application, and the scope of the application is defined by the appended claims and their equivalents.

Claims

1. A method of calculating available power transmission capacity of a power system, characterized by, The method comprises the following steps: obtaining operation data and grid topology data of a power system, and constructing an initial power flow model of the power system; constructing a photovoltaic power generation prediction model, the photovoltaic power generation prediction model being based on meteorological data and historical power generation data to predict photovoltaic injection power of a photovoltaic power station to the power system in a future period; the photovoltaic power generation prediction model comprises a dynamic health identification layer, a time-space disturbance identification layer, and a photovoltaic power generation prediction layer; the dynamic health identification layer identifies an aging influence parameter based on a factory use life of a photovoltaic power generation device and historical power generation data, and identifies a dust accumulation influence parameter of the photovoltaic power generation device based on rainfall data, atmospheric suspended particulate matter concentration, and meteorological data; the time-space disturbance identification layer extracts shape features, thickness features, speed features, and direction features of a cloud layer through a convolutional neural network to identify time-space interference parameters of the cloud layer on photovoltaic power generation, the time-space interference parameters comprising interference time data and time-space interference factors; the photovoltaic power generation prediction layer predicts the photovoltaic injection power based on normal direct radiation, horizontal scattered radiation, ambient temperature, and wind speed in the meteorological data, in combination with the aging influence parameter, the dust accumulation influence parameter, and the time-space interference parameters, to obtain the photovoltaic injection power; based on the operation data and the grid topology data, a transmission path of the photovoltaic injection power in the power system is identified, and a dynamic path loss is calculated based on the transmission path; the initial power flow model is updated based on the photovoltaic injection power and the dynamic path loss to obtain a predicted power flow model; under preset safe and stable constraint conditions, a maximum power transmission limit of the predicted power flow model is solved to determine an available power transmission capacity of the power system.

2. The method according to claim 1, wherein: the grid topology data of the power system is obtained through a power network management system; a power flow distribution path of the power generated by the photovoltaic power station to each part of the power grid is analyzed as a transmission path based on the grid topology data by using a power flow tracking algorithm; voltage phase angle data and voltage amplitude data collected by a phasor measurement unit deployed at a grid node are obtained; the line impedance parameters of the transmission path in the power flow tracking algorithm are corrected using the voltage phase angle data and the voltage amplitude data, and a dynamic path loss is calculated based on the corrected line impedance parameters.

3. The method according to claim 2, wherein: the process of correcting the line impedance parameters comprises: locking a line set that has a significant impact on photovoltaic power transmission based on the transmission path determined by the power flow tracking algorithm; extracting voltage phase angle data and voltage amplitude data of the first and last ends of the line set and adjacent key nodes; constructing an objective function of power flow equation residual minimization with the line impedance as a variable to be solved; solving the minimized objective function by using a Gauss-Newton method to obtain a correction value of the line impedance.

4. The method according to claim 1, wherein: A multi-dimensional operating state space containing photovoltaic injection power and key section active power is constructed through offline simulation calculation, and a multi-dimensional safety domain super body satisfying N-1 safety criteria, thermal stability constraints and voltage stability constraints is determined; The photovoltaic injection power and the key section active power are projected into the multi-dimensional operating state space to obtain a current operating point; Starting from the current operating point, the intersection point of the multi-dimensional safety domain super body boundary is calculated along the direction of the key section active power dimension coordinate axis, and the active power coordinate value corresponding to the intersection point is determined as the total power transmission capacity; The total power transmission capacity is reduced by the active power value of the power transmission section corresponding to the current operating point, and further reduced by a preset power transmission reliability margin and a capacity benefit margin to obtain the available power transmission capacity of the power system.

5. The power system available power transmission capacity calculation method according to claim 4, characterized in that: The key section is the transmission channel with the strongest photovoltaic injection power transmission capacity limitation obtained through multiple power flow calculations and static security analyses of the power system.

6. A power system available transfer capability calculation system characterized by comprising: It includes: A system construction module obtains operating data and grid topology structure data of the power system to construct an initial power flow model of the power system; A power calculation module constructs a photovoltaic power generation prediction model, which is based on meteorological data and historical power generation data to predict the photovoltaic injection power of the photovoltaic power station to the power system in a future period; The photovoltaic power generation prediction model includes a dynamic health identification layer, a time-space disturbance identification layer and a photovoltaic power generation prediction layer; The dynamic health identification layer identifies aging influence parameters based on the service life of photovoltaic power generation devices and historical power generation data, and identifies dust accumulation influence parameters of photovoltaic power generation devices based on rainfall data, atmospheric suspended particulate matter concentration and meteorological data; The time-space disturbance identification layer extracts the shape feature, thickness feature, speed feature and direction feature of the cloud layer through a convolutional neural network to identify time-space interference parameters of the cloud layer on photovoltaic power generation, which include interference time data and time-space interference factors; The photovoltaic power generation prediction layer predicts the photovoltaic injection power based on normal direct radiation, horizontal scattering radiation, ambient temperature and wind speed in the meteorological data, combined with the aging influence parameters, dust accumulation influence parameters and time-space interference parameters of the photovoltaic power generation devices; A loss identification module identifies the transmission path of the photovoltaic injection power in the power system based on the operating data and the grid topology structure data, and calculates the dynamic path loss based on the transmission path; A model updating module updates the initial power flow model based on the photovoltaic injection power and the dynamic path loss to obtain a predicted power flow model; A power transmission identification module determines the available power transmission capacity of the power system by solving the maximum power transmission limit of the predicted power flow model under preset safety and stability constraints.

7. A storage medium characterized by: The storage medium stores a computer program, which is executed by the processor to implement the power system available power transmission capacity calculation method of any one of claims 1 to 5.

Images